? A New Frontier Martin Kay Stanford University and The - - PowerPoint PPT Presentation

▶
a new frontier martin kay stanford university and the
SMART_READER_LITE
LIVE PREVIEW

? A New Frontier Martin Kay Stanford University and The - - PowerPoint PPT Presentation

Nov 16, 2023 568 likes •776 views

The European Union Translation ? A New Frontier Martin Kay Stanford University and The University of the Saarland Martin Kay Machine Translation 1 Martin Kay Machine Translation 2 T h The European Union e c o s t o f t r a

slide-1
SLIDE 1

Martin Kay Machine Translation

Martin Kay Stanford University and The University of the Saarland

Translation

A New Frontier

?

1 Martin Kay Machine Translation

The European Union

2 Martin Kay Machine Translation

The European Union

Danish Dutch English Finnish French German Greek Italian Portuguese Spanish Swedish Bulgarian Czech Estonian Hungarian Irish Latvian Lithuanian Maltese Polish Romanian Slovene Slovak

23 languages 2,500 (12.5%) of 20,000 staff 1% of the annual budget 40% of administration costs.

3

23 2

( ) = 253

T h e c

  • s

t

  • f

t r a n s l a t i

  • n

, i n t e r p r e t a t i

  • n

, p u b l i c a t i

  • n

, a n d l e g a l s e r v i c e s i n v

  • l

v e d i n m a k i n g I r i s h a n

  • f

f i c i a l E U l a n g u a g e i s e s t i m a t e d a t j u s t u n d e r € 3 . 5 m i l l i

  • n

a y e a r .

Martin Kay Machine Translation

300 authors and illustrators 800 English pages per day Translation into 14 languages

Maintenance Manuals Operation and Troubleshooting Guides Disassembly and Specifications Manuals Assembly Manuals Testing and Special Instructions Adjustment Guides Systems Operation Bulletins

4

slide-2
SLIDE 2

Martin Kay Machine Translation

Sound Meaning Language Source Target

5 Martin Kay Machine Translation

Weather reports Belles Lettres Advertising Manuals Scientific Papers Source Difficulty Easy Hard Target Quality High Low Dissemination Informative Assimilation Indicative There is a lot of stuff in this corner But this is where the action is!

6 Martin Kay Machine Translation

What is Translation?

  • A text that is based on a text in another

language and which

—has the same meaning —conveys the same information —has the same effect on its readers —gives the gist of the original —explains the original

  • It depends what you want
  • Generally a mixture of several of these

7 Martin Kay Machine Translation

Generally Speaking—

Foreign reader's experience should be as close as possible to that of a domestic reader. They should pass through parallel mental states.

8

N

  • m

e n t i

  • n
  • f

m e a n i n g !

slide-3
SLIDE 3

Martin Kay Machine Translation

Reference

9 Martin Kay Blood from a Stone 10

He sat down

Martin Kay Blood from a Stone 11

Language Situation 1 Situation 2 Update semantics

Martin Kay Blood from a Stone 12

Dateline: Damascus Athens Moscow Situation Updater

slide-4
SLIDE 4

Martin Kay Blood from a Stone 13

Problem 1: Keep track of the things

Reference

Martin Kay Translation

There are a lot of things

She took his wallet. — I saw it in her hand — I saw her do it — I could not believe it

14 Martin Kay Translation

but only a few things —

  • are on stage
  • are in the spotlight
  • preserve the momentum of what is being

said.

15 Martin Kay Machine Translation 16

She took his wallet.

— I saw it in her hand the thing — I saw her do it that the deed — I could not believe it what she did

slide-5
SLIDE 5

Martin Kay Machine Translation 17

She took his wallet.

— I saw it It was in her hand She was holding it — I saw her do it I saw her take it She took it as I watched She did it before my very eyes — I could not believe it I did not believe she could do such a thing

Martin Kay Machine Translation 18

She took his wallet.

— I saw her holding it — I saw it happen — I could not believe she could do it — It was unbelievable (to me) These do not mean the same thing out of context They do mean the same thing in context

Martin Kay Machine Translation

The sense of a word tells you what properties a thing must have for that word to refer to it. Things can be referred to by words with different senses.

Maybe choose the one that gives the most fluent result

19 Martin Kay Machine Translation

Types have senses Tokens have referents

20

slide-6
SLIDE 6

Martin Kay Machine Translation

Syntactic Processing

  • Find words with similar senses
  • Put them in a fluent order
  • Translation in Language teaching
  • Popular perception of translation
  • Machine translation

21 Martin Kay Machine Translation

Syntactic Processing

  • Find words with similar senses
  • Put them in a fluent order
  • St. Jerome: Vulgate (4th Century)
  • 20th Century
  • Translator is anonymous

22 Martin Kay Machine Translation

  • Words often have several senses.
  • You often cannot find words with similar

senses.

  • Even when you can, they often cannot be
  • rdered fluently.

23 Martin Kay Machine Translation

Trouble is

Different grammatical structures give different meanings.

24

The dog saw the cat The cat saw the dog Der Hund sah die Katze Den Hund sah die Katze

slide-7
SLIDE 7

Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Assurez-vous que vous n'avez rien

  • ublié dans le train

25 Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Assurez-vous que vous n'avez rien

  • ublié dans le train

26 Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Assurez-vous que vous n'avez rien

  • ublié dans le train

27 Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Assurez-vous que vous n'avez rien

  • ublié dans le train

28

slide-8
SLIDE 8

Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Assurez-vous que vous n'avez rien

  • ublié dans le train

29 Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Assurez-vous que vous n'avez rien

  • ublié dans le train

30 Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Assurez-vous que vous n'avez rien

  • ublié dans le train

31 Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Assurez-vous que vous n'avez rien

  • ublié dans le train

32

slide-9
SLIDE 9

Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Assurez-vous que vous n'avez rien

  • ublié dans le train

33 Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Assurez-vous que vous n'avez rien

  • ublié dans le train

34 Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Assurez-vous que vous n'avez rien

  • ublié dans le train

35 Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Make sure that you have forgotten nothing on the train Make sure that you have left nothing on the train Assurez-vous que vous n'avez rien

  • ublié dans le train

36

slide-10
SLIDE 10

Martin Kay Machine Translation

Versichern Sie Sich daß Sie nichts in dem Zug vergeßen haben Assurez-vous que vous n'avez rien

  • ublié dans le train

37

Make sure that you take all your belongings with you

Martin Kay Machine Translation

Semantic Processing

Replace words and phrases by words and phrases that mean the same Reorder

Pragmatic Processing

Replace words and phrases by words and phrases that will effect the same changes in the scene

38 Martin Kay Machine Translation

Pragmatic translation often involves addition or removal of substantive information from what is explicitly expressed.

39 Martin Kay Translation 40

Est_ce que ce train va a Perpignan?

slide-11
SLIDE 11

Martin Kay Translation 41

Does this train go to Perpignan? No, it stops in Beziers. Fährt dieser Zug nach Perpignan? Nein, er in Béziers endet hält

Martin Kay Machine Translation

Est-ce que c’est ta cousine? Non, je n’ai pas de cousine.

Is that your cousin?

  • No. I don’t have a cousin.

Is that woman your cousin?

42

female female

Martin Kay Machine Translation

Is that your cousin?

  • No. I don’t have a cousin.

Is that woman your cousin? girl

Est-ce que c’est ta cousine? Non, je n’ai pas de cousine.

43 Martin Kay Machine Translation

Abstraction

44

slide-12
SLIDE 12

Martin Kay Putting Linguistics back into Computational Linguistics

Phoneme Morpheme Sememe (?) Phrase (Distinctive) feature

Structural Linguistics

45

Linguistic Chemistry Elements

Martin Kay Putting Linguistics back into Computational Linguistics

Dimensions

46

front front back high low rounded unrounded

Martin Kay Putting Linguistics back into Computational Linguistics

Phonology

47

lætər → læɾər North American Flap

Martin Kay Putting Linguistics back into Computational Linguistics

Generalization

One statement covers many instances.

48

slide-13
SLIDE 13

Martin Kay Putting Linguistics back into Computational Linguistics

Morphographemics

define sib [j | s | x | z | s h | c h] ; define consonant [ b | c | d | f | g | h | j | k | l | m | n | p | q | r | s | t | v | w | x | y | z ] ; define vowel [a | e | i | o | u] ; define boundary [.#. | % +]; define optional [ %? (->) 0] ; define YtoIE [ y -> i e || consonant _ EM alpha] ; define IEtoY [ i e -> y || _ EM i ] ; define Edeletion1 [ e -> 0 || vowel consonant _ EM vowel ] ; define Edeletion2 [ e EM e -> EM e ] ; define Einsertion [ [..] -> e || [sib | o] (diacritic) EM _ s EM ] ; define gemination [ b -> b b, c -> c k, d -> d d, f -> f f, g -> g g, l -> l l, m -> m m, n -> n n, p -> p p, r -> r r, s -> s s, t -> t t || vowel _ EM vowel ] ; define DiacriticDeletion [ diacritic -> 0 ] ; define BoundaryDeletion [ [BM | EM] -> 0] ;

49

103 States 2368 Transitions

Martin Kay Putting Linguistics back into Computational Linguistics

Morphology

50

Case N u m b e r Gender Declensions Conjugations

Martin Kay Putting Linguistics back into Computational Linguistics 51

stem cat verb conjugation 1 lex parler

parl e

suffix cat verb conjugation 1 tense pres person 1 | 3 number sing word cat verb lex ?Lex tense ?Tense person ?Person number ?Number stem cat verb conjugation ?Conj lex ?Lex suffix cat verb conjugation ?Conj tense ?Tense person ?Person number ?Number

⇒

word cat verb lex parler tense pres person 1 | 3 number sing

parle One rule for all finite verbs

Martin Kay Putting Linguistics back into Computational Linguistics

Syntax

Agreement Head features Subcategorization Modification Long-distance dependencies

52

slide-14
SLIDE 14

Syntax Decrease the number

  • f types of entity

Martin Kay Machine Translation

Language Abstraction

Phonology / Orthography Morphology Semantics

53 Martin Kay Machine Translation

Phonology / Orthography Morphology Syntax Semantics tying = tie+ing loved = love + ed

Language Abstraction

54 Martin Kay Machine Translation

Phonology / Orthography Morphology Syntax Semantics loves = love+pl, love+3sg children = child+pl

Language Abstraction

55 Martin Kay Machine Translation

Phonology / Orthography Morphology Syntax Semantics The dog chased the cat = the cat was chased by the dog

Language Abstraction

56

slide-15
SLIDE 15

Martin Kay Machine Translation

Phonology / Orthography Morphology Syntax Semantics Brutus killed Ceasar => Caesar is dead

Language Abstraction

57 Martin Kay Machine Translation

Language Abstraction

Phonology / Orthography Morphology Syntax Semantics Abstraction Orthogonal masc, fem, neut sing, plur nom, acc, gen, dat past, present, future perfective, imperfective ...

58 Martin Kay Machine Translation

Language Abstraction

Phonology / Orthography Morphology Syntax Semantics Abstracti Orthogonal active, passive decl, inter, imper main, subordinate subj, obj, mod pos, neg ...

59 Martin Kay Machine Translation

Vauquois Triangle

Phonology / Orthography Morphology Syntax Semantics The more abstract the description

60

slide-16
SLIDE 16

Martin Kay Machine Translation

Vauquois Triangle

Phonology / Orthography Morphology Syntax Semantics

Interlingua

The more similar the languages

61 Martin Kay Machine Translation

Vauquois Triangle

Phonology / Orthography Morphology Syntax Semantics Interlingua Transfer Analysis Synthesis

62 Martin Kay Machine Translation

Vauquois Triangle

Phonology / Orthography Morphology Syntax Semantics Interlingua Transfer Analysis Synthesis

63 Martin Kay Machine Translation

Language The World

L'Arbitraire du Signe

64

slide-17
SLIDE 17

Martin Kay Machine Translation

Transfer Analysis Synthesis

65 Martin Kay Machine Translation

L'Arbitraire du Signe

I left nothing on the train I had all my stuff with me when I got off the train

66 Martin Kay Machine Translation

Vauquois Triangle

Syntax

Semantic Processing

67 Martin Kay Machine Translation

L'Arbitraire du Signe

Pragmatic Processing

68

slide-18
SLIDE 18

Martin Kay Machine Translation

Pragmatic Processing

69 Martin Kay Machine Translation

People, see everything —in a context —from a point of view

70

Situated Embodied